A covariance regression model

2011/02/28 by Hoff, Peter D., Niu, Xiaoyue · 2 citations
#62H25 #FOS: Computer and information sciences #Methodology (stat.ME)

paper · doi:10.48550/arxiv.1102.5721

Abstract

Classical regression analysis relates the expectation of a response variable to a linear combination of explanatory variables. In this article, we propose a covariance regression model that parameterizes the covariance matrix of a multivariate response vector as a parsimonious quadratic function of explanatory variables. The approach is analogous to the mean regression model, and is similar to a factor analysis model in which the factor loadings depend on the explanatory variables. Using a random-effects representation, parameter estimation for the model is straightforward using either an EM-algorithm or an MCMC approximation via Gibbs sampling. The proposed methodology provides a simple but flexible representation of heteroscedasticity across the levels of an explanatory variable, improves estimation of the mean function and gives better calibrated prediction regions when compared to a homoscedastic model.

Cited by

• Generalized Correlation Regression for Disentangling Dependence in Clustered Data
• Test of partial effects for Frechet regression on Bures-Wasserstein manifolds

Related

• Fixed and Random Covariance Regression Analyses
• Bayesian Nonparametric Covariance Regression
• Reducing Subspace Models for Large-Scale Covariance Regression
• Mixed-type multivariate response regression with covariance estimation
• A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity